Base64 Encode and Decode Text

This deterministic converter handles Unicode as UTF-8, supports both standard RFC 4648 Base64 and its URL-safe variant, and rejects malformed decode input instead of returning corrupted text. Use it interactively in the browser or call the API when an application needs repeatable conversion. Each successful response contains one result string, making the output straightforward to copy, store, compare, or pass to the next step in a pipeline.

Encode UTF-8 text into a portable Base64 representation

Choose encode mode when you have ordinary text and need an ASCII representation suitable for a text-only field, configuration value, JSON document, or integration boundary. The converter first represents the supplied string as UTF-8 bytes, so accented letters, emoji, non-Latin scripts, line breaks, and punctuation are processed consistently rather than being truncated to an older single-byte character set. It then groups those bytes according to RFC 4648 and returns the Base64 result. With the standard variant, the output uses letters, digits, plus, and slash, with canonical equals-sign padding when the final byte group is incomplete. With the URL-safe flag enabled, plus and slash become hyphen and underscore, and generated padding is omitted to produce a compact value suitable for route segments or query parameters. Base64 is an encoding rather than encryption: anyone with the value can reverse it, so confidential material still needs appropriate transport and storage protection. The returned object contains only the result string, which keeps scripted processing simple and deterministic.

Decode strictly instead of accepting corrupted input

Choose decode mode when the supplied text is already Base64 and you want its original UTF-8 text. The selected variant matters during validation. Standard mode accepts the standard plus and slash symbols, while URL-safe mode accepts hyphen and underscore; characters from the other alphabet are treated as invalid rather than silently rewritten. The decoder also checks that padding appears only at the end, that no more than two padding characters are present, that explicit padding matches the encoded length, and that the length can describe complete bytes. It verifies unused trailing bits too, which prevents multiple noncanonical strings from being treated as the same byte sequence. Whitespace is rejected because the input contract calls for a Base64 string, not MIME-formatted lines with formatting characters. Finally, decoded bytes must form valid UTF-8. A binary file may be perfectly valid Base64 yet still be unsuitable for this text-focused capability; in that case, decoding returns an input error rather than replacement characters. These checks make failures visible at the integration boundary, where they are much easier to diagnose than damaged content discovered later.

Select standard or URL-safe Base64 for the receiving system

The correct variant is determined by the system that will consume the value. Standard Base64 is common in encoded document fields, data interchange, and specifications that explicitly show plus, slash, and equals signs. URL-safe Base64 is designed for contexts where plus and slash have special meaning, including URLs, filenames, cookies, and token components. Setting url_safe changes both encoding and decoding, so it should reflect the alphabet actually expected by the other side rather than being used as a cleanup switch. When decoding URL-safe values, either canonical padding or omitted padding is accepted when the length permits it; when encoding, padding is deliberately omitted for the compact form widely used in URL contexts. For dependable pipelines, store the variant alongside the data or define it in the surrounding protocol instead of guessing from a sample that may contain only letters and digits. The browser tool is useful for one-off inspection, while the API costs $0.002 per request for automated workflows. Both paths use the same pure conversion logic, produce the same result, make no network call from the algorithm, and do not rely on random values or the current time.